test: connect inference/eval tests to algorithms

[begumcig]

Description

Adding simple evaluation runs to the algorithm tests to check if model is capable of running inference, and the inference outputs are compatible with the evaluation agent.
Small bug fixes here and there to get the tests up and running

Type of Change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[johannaSommer]

I cant believe you did this and its almost flawless on the first take, i really appreciate the effort you put into this, it's so key! just a bit of comments and logic changes then we're already GTG

[johannaSommer]

@nifleisch @johnrachwan123 if we require the fullgraph this often, should we consider making this the new default?

[nifleisch]

Agreed. fullgraph=False is also the default torch.compile uses so it would be more convenient to match this default. Especially since fullgraph=True can lead to graph breaks for certain models and is less stable.

[johannaSommer]

🚔

[nifleisch]

I like the PR, approved! 🙌

[nifleisch]

Agreed. fullgraph=False is also the default torch.compile uses so it would be more convenient to match this default. Especially since fullgraph=True can lead to graph breaks for certain models and is less stable.

[johannaSommer]

LGTM! 🫡